CN117992859B - Early-stage fault early-warning and identifying method and device for electromechanical equipment provided with SCADA system - Google Patents
Early-stage fault early-warning and identifying method and device for electromechanical equipment provided with SCADA system Download PDFInfo
- Publication number
- CN117992859B CN117992859B CN202410400068.2A CN202410400068A CN117992859B CN 117992859 B CN117992859 B CN 117992859B CN 202410400068 A CN202410400068 A CN 202410400068A CN 117992859 B CN117992859 B CN 117992859B
- Authority
- CN
- China
- Prior art keywords
- matrix
- data matrix
- dynamic network
- time window
- covariance matrix
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 67
- 239000011159 matrix material Substances 0.000 claims abstract description 247
- 238000005259 measurement Methods 0.000 claims abstract description 65
- 238000005070 sampling Methods 0.000 claims abstract description 49
- 238000011156 evaluation Methods 0.000 claims abstract description 48
- 239000013598 vector Substances 0.000 claims abstract description 43
- 238000013139 quantization Methods 0.000 claims abstract description 37
- 230000036541 health Effects 0.000 claims abstract description 31
- 238000000354 decomposition reaction Methods 0.000 claims abstract description 27
- 238000012216 screening Methods 0.000 claims abstract description 12
- 238000004364 calculation method Methods 0.000 claims abstract description 6
- 238000013507 mapping Methods 0.000 claims abstract description 6
- 238000012544 monitoring process Methods 0.000 claims description 18
- 238000004590 computer program Methods 0.000 claims description 10
- 230000008878 coupling Effects 0.000 claims description 9
- 238000010168 coupling process Methods 0.000 claims description 9
- 238000005859 coupling reaction Methods 0.000 claims description 9
- 238000013528 artificial neural network Methods 0.000 claims description 8
- 230000004044 response Effects 0.000 claims description 8
- 230000007704 transition Effects 0.000 claims description 6
- 238000004458 analytical method Methods 0.000 claims description 5
- 238000010276 construction Methods 0.000 claims description 3
- 238000009434 installation Methods 0.000 claims description 2
- 230000007787 long-term memory Effects 0.000 claims description 2
- 230000006403 short-term memory Effects 0.000 claims description 2
- 238000012545 processing Methods 0.000 abstract description 11
- 230000015654 memory Effects 0.000 description 10
- 238000010586 diagram Methods 0.000 description 8
- 230000008859 change Effects 0.000 description 7
- 238000004891 communication Methods 0.000 description 6
- 230000006870 function Effects 0.000 description 6
- 238000012423 maintenance Methods 0.000 description 5
- 239000003550 marker Substances 0.000 description 5
- 230000008569 process Effects 0.000 description 5
- 238000012360 testing method Methods 0.000 description 5
- 230000007547 defect Effects 0.000 description 4
- 230000003287 optical effect Effects 0.000 description 4
- 230000000694 effects Effects 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 230000007613 environmental effect Effects 0.000 description 3
- 238000004519 manufacturing process Methods 0.000 description 3
- 230000006399 behavior Effects 0.000 description 2
- 238000004422 calculation algorithm Methods 0.000 description 2
- 238000003745 diagnosis Methods 0.000 description 2
- 239000000835 fiber Substances 0.000 description 2
- 230000007246 mechanism Effects 0.000 description 2
- 230000000644 propagated effect Effects 0.000 description 2
- 239000004065 semiconductor Substances 0.000 description 2
- 238000012549 training Methods 0.000 description 2
- 230000009471 action Effects 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 230000015556 catabolic process Effects 0.000 description 1
- 238000006731 degradation reaction Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000007667 floating Methods 0.000 description 1
- 239000004973 liquid crystal related substance Substances 0.000 description 1
- 238000013178 mathematical model Methods 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 230000002035 prolonged effect Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/16—Matrix or vector computation, e.g. matrix-matrix or matrix-vector multiplication, matrix factorization
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/211—Selection of the most significant subset of features
- G06F18/2113—Selection of the most significant subset of features by ranking or filtering the set of features, e.g. using a measure of variance or of feature cross-correlation
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Mathematical Physics (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Analysis (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Computation (AREA)
- Mathematical Optimization (AREA)
- Evolutionary Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computational Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Pure & Applied Mathematics (AREA)
- Computing Systems (AREA)
- Software Systems (AREA)
- Databases & Information Systems (AREA)
- Algebra (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses an early-stage fault early-warning and identifying method and device for electromechanical equipment provided with a SCADA system, which relate to the field of data processing and comprise the following steps: mapping the variables selected by the SCADA system into nodes of a dynamic network, constructing a corresponding health state theoretical prediction model, inputting an actual measurement data matrix of a previous sampling period into the health state theoretical prediction model, and obtaining a reference data matrix of a current sampling period; the method comprises the steps of respectively constructing covariance matrixes of actual measurement data matrixes and covariance matrixes of reference data matrixes, respectively carrying out feature decomposition, screening out key nodes of a dynamic network according to feature vectors corresponding to maximum feature values obtained by decomposition and element ranks of the feature vectors, respectively calculating quantization indexes of the dynamic network of the actual measurement data matrixes and the reference data matrixes in each time window, calculating system evaluation indexes, generating early warning signals and determining fault types based on the system evaluation indexes and the key nodes corresponding to the system evaluation indexes, and solving the problems of high calculation cost and poor generalization capability.
Description
Technical Field
The invention relates to the field of data processing, in particular to an early-stage fault early-warning and identifying method and device for electromechanical equipment provided with a SCADA system.
Background
With the deep development of industrial automation, the demand for stable and efficient operation of complex electromechanical equipment is increasing. These devices are typically composed of a variety of mechanical, electrical and control components that form a multi-dimensional time-varying complex system. In these systems, there is a complex mutual coupling relationship between the components or subsystems, and failure of any one component may affect the normal operation of the entire system, and serious failure may even have a significant impact on the stability of the entire production process or service. Therefore, for the complex electromechanical equipment, the effective fault early warning technology can reduce operation and maintenance cost and reduce the requirement of emergency maintenance, and the service life of the equipment is prolonged through effective predictive maintenance, so that the production safety and efficiency are ensured. Therefore, the method has extremely important research and application value for implementing the fault early warning technology on the complex electromechanical equipment.
At present, most of complex electromechanical equipment is provided with a data acquisition and monitoring control system (Supervisory Control And Data Acquisition, namely an SCADA system), and early defect early warning and identification of the complex electromechanical equipment based on the SCADA system becomes an effective means. Defect early warning methods for complex electromechanical equipment equipped with SCADA systems are mainly divided into two main categories: a mechanism-based model and a data-driven model-based method. Because of the complex mapping relationship between the fault types of the internal physical components of the devices and the externally observable fault characteristic quantities, the traditional early warning method based on the mechanism model faces the problem that an accurate mathematical model is difficult to build. Meanwhile, the fault early warning method based on data driving also meets a plurality of challenges in practical application. For example, it is often difficult to obtain sufficiently large and representative fault sample data, which affects the training effect of the model. Furthermore, differences in manufacturing process, installation environment, age, and operating status of the different devices limit the robustness of models built based on historical data. Meanwhile, the artificial intelligence method requires a large amount of data and computing resources, which increases the cost and puts higher real-time operation requirements on hardware devices. The fault diagnosis and early warning method of the current complex electromechanical equipment also has the common problem: the equipment state monitoring data has the characteristics of global and diversified, and complex coupling relations exist among all the components, so that the method has remarkable effect on processing the faults of the specific components, but the method is worry about fault early warning of the whole equipment, and the problem of effectively fusing the multiple data becomes an important difficulty.
In order to describe critical phase change dynamic characteristics of a multivariable complex system, a concept of dynamic network markers (DYNAMICAL NETWORK MARKER, DNM) is proposed by a learner, and the effectiveness of the proposed method is demonstrated in the field of fault diagnosis of complex electromechanical equipment. Fang Ruiming et al first applied DNM in early defect early warning of wind turbines and achieved good results. However, the method requires a clustering algorithm or other heuristic programs to screen the key nodes of the network, and has high calculation cost and poor generalization capability.
Disclosure of Invention
The application aims to provide an early-stage fault early-warning and identifying method and device for electromechanical equipment provided with a SCADA system aiming at the technical problems.
In a first aspect, the present invention provides a method for early warning and identifying faults in electromechanical equipment equipped with a SCADA system, comprising the steps of:
Selecting variables closely related to the operation state and having complex coupling relation from all main monitoring items of an SCADA system of the electromechanical equipment, and mapping the variables into nodes of a dynamic network;
Constructing a health state theoretical prediction model aiming at each node of a dynamic network, acquiring actual measurement historical data of each node acquired by an SCADA system in each sampling time in a previous sampling period when electromechanical equipment is in a health state, constructing an actual measurement data matrix in the previous sampling period, and inputting the actual measurement data matrix in the previous sampling period into the health state theoretical prediction model to obtain a reference data matrix in the current sampling period;
respectively constructing covariance matrixes of measured data matrixes and covariance matrixes of reference data matrixes of each time window divided in a current sampling period, respectively performing feature decomposition, calculating feature values and corresponding feature vectors obtained after the decomposition, and screening key nodes of a dynamic network according to the feature vectors corresponding to the maximum feature values and element ranks thereof;
The method comprises the steps of respectively calculating quantization indexes of a dynamic network of an actual measurement data matrix and a reference data matrix in each time window based on key nodes of the dynamic network, calculating a system evaluation index in each time window according to the quantization indexes of the dynamic network of the actual measurement data matrix and the reference data matrix in each time window, generating early warning signals based on the system evaluation indexes and the key nodes of the corresponding dynamic network, and determining fault types.
Preferably, the theoretical state of health prediction model comprises a trained long and short term memory neural network.
Preferably, the actual measurement history data is normalized data.
Preferably, the covariance matrix of the measured data matrix is expressed as:
；
The covariance matrix of the reference data matrix is expressed as:
；
Wherein, AndThe elements in the measured data matrix A and the reference data matrix B are respectively represented, and the corresponding values are covariance of the ith and jth variables.
Preferably, the calculating and decomposing the feature value and the corresponding feature vector thereof, and screening out the key nodes of the dynamic network according to the feature vector corresponding to the maximum feature value and the element ranking thereof specifically comprises the following steps:
The covariance matrix of the measured data matrix and the covariance matrix of the reference data matrix can both represent variables Sample covariance matrix/>, consisting of N observationsAnd carrying out characteristic decomposition on the sample covariance matrix, wherein the characteristic decomposition is shown in the following formula:
；
Wherein, Is a diagonal matrix containing eigenvalues/>, of the sample covariance matrix C；Is an orthogonal matrix comprising eigenvectors/>, of a sample covariance matrix C，；
The elements of the ith row and jth column of the sample covariance matrix at time window t are represented as:
；
Wherein, ，Is the variable/>, in the sample covariance matrix CAndCovariance,Representing the maximum eigenvalue,AndVariable/>, respectivelyAndFeature vector/>, corresponding to maximum feature value, in sample covariance matrix of time window tAn ith element and a jth element; comparing the eigenvector/>, corresponding to the maximum eigenvalueSelecting variables corresponding to the first h elements of the numerical ranking as key nodes of the dynamic network, wherein the set of index marks is defined asThe value of h is set to satisfyAndWherein, the method comprises the steps of, wherein,Representing the variableThe ith element of the eigenvector corresponding to the largest eigenvalue in the sample covariance matrix of the critical transition time window t', the P value is a positive real number of [0,1 ]IsUpper bound of all element numbers in the whole system.
Preferably, the method includes calculating quantization indexes of the dynamic network in each time window based on the key nodes of the dynamic network, respectively, and calculating a system evaluation index in each time window according to the quantization indexes of the dynamic network in each time window of the actual measurement data matrix and the reference data matrix, specifically including:
Respectively calculating quantization indexes of the dynamic network of the actual measurement data matrix in a time window t by adopting the following steps of And a quantization index/>, with reference to the data matrix, of the dynamic network over a time window t：
；
；
Wherein,The element of the ith row and the jth column of the covariance matrix characteristic of the actual measurement data matrix at the time window t is h A, which is the first h A elements in the feature vector corresponding to the maximum feature value obtained after the covariance matrix characteristic of the actual measurement data matrix is decomposed, and the set of index marks of the key nodes of the corresponding dynamic network is，The element of the ith row and the jth column of the covariance matrix feature of the reference data matrix at the time window t is h B, which is the first h B elements in the feature vector corresponding to the maximum feature value obtained by decomposing the covariance matrix feature of the reference data matrix, and the set of index marks of the key nodes of the corresponding dynamic network is；
Calculating a system evaluation index of the time window t by adopting the following steps:
；
Wherein, Representing a system evaluation index; /(I)Covariance matrix use element set/>, which is measured data matrixCalculating quantization index of dynamic network in time window t, element setIs randomly selected from all elements in the eigenvector corresponding to the maximum eigenvalue of the covariance matrix of the measured data matrix, and the element setNumber and number of (2)The number of the sets is consistent,Covariance matrix usage element set/>, which is a reference data matrixCalculating quantization index of dynamic network in time window t, element setIs randomly selected from all elements in the eigenvector corresponding to the maximum eigenvalue of the covariance matrix of the reference data matrix, and the element setNumber andThe number of the sets is consistent.
Preferably, the method includes generating early warning signals and determining fault types based on system evaluation indexes and key nodes of corresponding dynamic networks, and specifically includes:
in response to determining that the system evaluation index is greater than the threshold, generating an early warning signal according to the corresponding time window Analyzing the key nodes corresponding to the index identifiers in the set to obtain corresponding fault types;
In response to determining that the system evaluation index is less than or equal to the threshold, the electromechanical equipment is determined to be in a healthy state.
In a second aspect, the present invention provides an early-stage fault early-warning and identifying device for electromechanical equipment equipped with a SCADA system, comprising:
the dynamic network construction module is configured to select variables which are closely related to the operation state and have complex coupling relations from all main monitoring items of the SCADA system of the electromechanical equipment, and map the variables into nodes of the dynamic network;
The system comprises a theoretical prediction module, a dynamic network and a dynamic network, wherein the theoretical prediction module is configured to construct a health state theoretical prediction model for each node of the dynamic network, acquire actual measurement historical data of each node acquired by an SCADA system in a previous sampling period at each sampling time when the electromechanical equipment is in a health state, construct an actual measurement data matrix of the previous sampling period, and input the actual measurement data matrix of the previous sampling period into the health state theoretical prediction model to acquire a reference data matrix of the current sampling period;
The key node screening module is configured to respectively construct a covariance matrix of the actual measurement data matrix and a covariance matrix of the reference data matrix of each time window divided in the current sampling period, respectively perform characteristic decomposition, calculate characteristic values obtained after the decomposition and corresponding characteristic vectors thereof, and screen key nodes of the dynamic network according to the characteristic vectors corresponding to the maximum characteristic values and element ranks thereof;
The fault analysis module is configured to calculate quantization indexes of the dynamic network of the actual measurement data matrix and the reference data matrix in each time window based on key nodes of the dynamic network respectively, calculate a system evaluation index in each time window according to the quantization indexes of the dynamic network of the actual measurement data matrix and the reference data matrix in each time window, generate early warning signals based on the system evaluation indexes and the key nodes of the corresponding dynamic network, and determine fault types.
In a third aspect, the present invention provides an electronic device comprising one or more processors; and storage means for storing one or more programs which, when executed by the one or more processors, cause the one or more processors to implement the method as described in any of the implementations of the first aspect.
In a fourth aspect, the present invention provides a computer readable storage medium having stored thereon a computer program which, when executed by a processor, implements a method as described in any of the implementations of the first aspect.
Compared with the prior art, the invention has the following beneficial effects:
(1) The early-fault early warning and identifying method for the electromechanical equipment provided with the SCADA system dynamically predicts the reference data matrix in a mode of predicting the health state theoretical prediction model based on the neural network pair, and compared with the traditional method that a fixed reference object is selected by a dynamic network marker, the early-fault early warning and identifying method for the electromechanical equipment provided with the SCADA system is more suitable for the electromechanical equipment which is operated in a complex environment with larger change, and has stronger robustness.
(2) The early-stage fault early-warning and identifying method for the electromechanical equipment provided with the SCADA system adopts the dynamic network marker method based on the sample covariance matrix, screens key nodes of the dynamic network according to the ranking of elements in the feature vector from the statistical perspective, can better understand the behavior change of the system, and solves the problems that the traditional dynamic network marker method needs more perfect priori knowledge for screening the key network, and the algorithm for screening the key nodes of the dynamic network needs special adjustment for different application scenes, has limited generalization capability and the like.
(3) The early-stage fault early-warning and identifying method for the electromechanical equipment provided with the SCADA system can identify the variables playing a key role in the system behavior by comparing the elements in different feature vectors, so that the key features of the dynamic network are determined, and the method can be used as a reference basis for further identifying the equipment fault reasons.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings that are needed in the description of the embodiments will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is an exemplary device frame pattern to which an embodiment of the present application may be applied;
FIG. 2 is a flow chart of an early failure warning and identifying method for an electromechanical device equipped with a SCADA system according to an embodiment of the present application;
FIG. 3 is a schematic diagram of a continuous monitoring project of an SCADA system of a complex electromechanical device selected in an early-stage fault early-warning and identifying method of the electromechanical device equipped with the SCADA system according to an embodiment of the present application;
FIG. 4 shows the calculation of the early failure warning and identifying method of the electromechanical device of the embodiment 1 of the present application under the condition of breeze in healthy state A result graph;
FIG. 5 shows the early failure warning and identifying method of the electromechanical device of embodiment 1 of the present application under strong wind conditions A result graph;
FIG. 6 is a diagram showing the early failure warning and identifying method of the electromechanical device of the embodiment 1 of the present application under the condition of heavy wind in a healthy state A result graph;
FIG. 7 shows an early failure warning and identifying method for an electromechanical device equipped with a SCADA system before and after failure in accordance with embodiment 2 of the present application A result graph;
FIG. 8 is a ranking chart of element values of feature vectors calculated by an early-stage fault early-warning and identifying method of electromechanical equipment provided with a SCADA system at a fault early-warning point of the electromechanical equipment according to embodiment 2 of the present application;
FIG. 9 is a schematic diagram of an early failure warning and identification device for an electromechanical device equipped with a SCADA system according to an embodiment of the present application;
Fig. 10 is a schematic structural view of a computer device suitable for use in implementing an embodiment of the present application.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention will be described in further detail below with reference to the accompanying drawings, and it is apparent that the described embodiments are only some embodiments of the present invention, not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
FIG. 1 illustrates an exemplary device architecture 100 in which the SCADA system-equipped early-stage fault warning and identification method or the SCADA system-equipped early-stage fault warning and identification device of an electromechanical device may be applied in accordance with embodiments of the present application.
As shown in fig. 1, the apparatus architecture 100 may include a first terminal device 101, a second terminal device 102, a third terminal device 103, a network 104, and a server 105. The network 104 is used as a medium to provide communication links between the first terminal device 101, the second terminal device 102, the third terminal device 103, and the server 105. The network 104 may include various connection types, such as wired, wireless communication links, or fiber optic cables, among others.
The user may interact with the server 105 via the network 104 using the first terminal device 101, the second terminal device 102, the third terminal device 103, to receive or send messages, etc. Various applications, such as a data processing class application, a file processing class application, and the like, may be installed on the terminal device one 101, the terminal device two 102, and the terminal device three 103.
The first terminal device 101, the second terminal device 102 and the third terminal device 103 may be hardware or software. When the first terminal device 101, the second terminal device 102, and the third terminal device 103 are hardware, they may be various electronic devices, including but not limited to smart phones, tablet computers, laptop computers, desktop computers, and the like. When the first terminal apparatus 101, the second terminal apparatus 102, and the third terminal apparatus 103 are software, they can be installed in the above-listed electronic apparatuses. Which may be implemented as multiple software or software modules (e.g., software or software modules for providing distributed services) or as a single software or software module. The present invention is not particularly limited herein.
The server 105 may be a server that provides various services, such as a background data processing server that processes files or data uploaded by the terminal device one 101, the terminal device two 102, and the terminal device three 103. The background data processing server can process the acquired file or data to generate a processing result.
It should be noted that, the early-stage fault early-warning and identifying method for the electromechanical equipment provided with the SCADA system provided by the embodiment of the application may be executed by the server 105, or may be executed by the first terminal device 101, the second terminal device 102, or the third terminal device 103, and correspondingly, the early-stage fault early-warning and identifying device for the electromechanical equipment provided with the SCADA system may be provided in the server 105, or may be provided in the first terminal device 101, the second terminal device 102, or the third terminal device 103.
It should be understood that the number of terminal devices, networks and servers in fig. 1 is merely illustrative. There may be any number of terminal devices, networks, and servers, as desired for implementation. In the case where the processed data does not need to be acquired from a remote location, the above-described apparatus architecture may not include a network, but only a server or terminal device.
Fig. 2 shows an early-stage fault early-warning and identifying method for electromechanical equipment equipped with a SCADA system according to an embodiment of the present application, including the following steps:
s1, selecting variables which are closely related to the operation state and have complex coupling relation from all main monitoring items of an SCADA system of the electromechanical equipment, and mapping the variables into nodes of a dynamic network.
Specifically, the electromechanical device of the embodiment of the application is a wind turbine generator set with a model GE1.5SLE provided with an SCADA system, the rated capacity of the wind turbine generator set is 1.5MW, and the sampling frequency of the SCADA system is 1 time/minute. The monitoring range of the SCADA system covers 47 continuous monitoring projects, which are shown in fig. 3, and comprise 4 weather information items, 38 running condition information items of subsystems or components of the wind turbine generator, and various signal types such as angles, pressures, temperatures, speeds (including rotating speeds and wind speeds), vibration and electrical factors and the like. The embodiment of the application is simulated by taking MATLAB as a working platform. The embodiment of the application selects 29 variables with strong correlation as sample data of the dynamic network of the embodiment of the application, and the selected item names are shown in table 1.
29 SCADA continuous monitoring items selected in Table 1;
S2, constructing a health state theoretical prediction model aiming at each node of the dynamic network, acquiring actual measurement historical data of each node acquired by the SCADA system in each sampling time in the previous sampling period when the electromechanical equipment is in a health state, constructing an actual measurement data matrix of the previous sampling period, and inputting the actual measurement data matrix of the previous sampling period into the health state theoretical prediction model to obtain a reference data matrix of the current sampling period.
In a specific embodiment, the theoretical state of health prediction model includes a trained long-short term memory neural network.
In a specific embodiment, the measured history data is normalized data.
Specifically, the historical sample data acquired by the SCADA system when each node of the electromechanical equipment is in a healthy state is utilized to normalize the historical sample data, and the long-short-time memory neural network is trained by adopting the normalized historical sample data, so that the trained long-short-time memory neural network is obtained, namely, the healthy state theoretical prediction model. The long-short-term memory neural network belongs to the existing network model, and the specific structure thereof is not described herein. And applying a long-short-time memory neural network with regression prediction capability to historical sample data of the selected SCADA system when the monitored item is in a healthy state, sequentially training a healthy state theoretical prediction model of each node of the dynamic network, and inputting an actual measurement data matrix of a previous sampling period into the trained healthy state theoretical prediction model to obtain a predicted reference data matrix of the current sampling period. The reference data matrix thus constructed and the measured data matrix may correspond to each other in time series.
The initial time point corresponding to the actually measured data matrix in the current sampling period is t 0, and the total number of time points is n. Taking an actual measurement data matrix of a previous sampling period (comprising m time points, m being greater than or equal to n) before a starting time point t 0 as an input, predicting by a health state theory prediction model to obtain a reference data matrix, wherein elements in the reference data matrix are data of the future (n time points). The reference data matrix and the measured data matrix in the current sampling period of length n from the start time point t 0 in time series may thus correspond to each other.
S3, respectively constructing covariance matrixes of the actual measurement data matrixes and covariance matrixes of the reference data matrixes of each time window divided in the current sampling period, respectively performing feature decomposition, calculating feature values obtained after the decomposition and feature vectors corresponding to the feature values, and screening key nodes of the dynamic network according to the feature vectors corresponding to the maximum feature values and element ranks of the feature vectors.
In a specific embodiment, the covariance matrix of the measured data matrix is expressed as:
；
The covariance matrix of the reference data matrix is expressed as:
；
Wherein, AndThe elements in the measured data matrix A and the reference data matrix B are respectively represented, and the corresponding values are covariance of the ith and jth variables.
In a specific embodiment, calculating the feature value obtained after decomposition and the feature vector corresponding to the feature value, and screening out key nodes of the dynamic network according to the feature vector corresponding to the maximum feature value and element ranking thereof, wherein the method specifically comprises the following steps:
The covariance matrix of the measured data matrix and the covariance matrix of the reference data matrix can both represent variables Sample covariance matrix/>, consisting of N valuesAnd carrying out characteristic decomposition on the sample covariance matrix, wherein the characteristic decomposition is shown in the following formula:
；
Wherein, Is a diagonal matrix containing eigenvalues/>, of the sample covariance matrix C；Is an orthogonal matrix comprising eigenvectors/>, of a sample covariance matrix C，；
The elements of the ith row and jth column of the sample covariance matrix at time window t are represented as:
；
Wherein, ，Is the variable/>, in the sample covariance matrix CAndCovariance,Representing the maximum eigenvalue,AndVariable/>, respectivelyAndFeature vector/>, corresponding to maximum feature value, in sample covariance matrix of time window tAn ith element and a jth element; comparing the eigenvector/>, corresponding to the maximum eigenvalueSelecting variables corresponding to the first h elements of the numerical ranking as key nodes of the dynamic network, wherein the set of index marks is defined asThe value of h is set to satisfyAndWherein, the method comprises the steps of, wherein,Representing the variableThe ith element of the eigenvector corresponding to the largest eigenvalue in the sample covariance matrix of the critical transition time window t', the P value is a positive real number of [0,1 ]IsUpper bound of all element numbers in the whole system.
Specifically, the filtered measured data matrix A and the predicted reference data matrix B have the same structure, i.eThe data in the actual measurement data matrix A and the reference data matrix B are time sequence sample data, N corresponds to the number of time points of the time sequence sample data, and N represents the dimension of a variable. In order to analyze the change trend of the time sequence sample data more accurately, proper K is selected, the time sequence sample data is divided into N/K independent time windows, and covariance matrixes of the actual measurement data matrix A and the reference data matrix B of each time window are calculated.
The following is an explanation about the covariance matrix:
for two random variables X and Y, their covariance is defined as:
；
Wherein, AndIs the observed value of each variable,AndAre their average values, and M is the number of samples.
Covariance measures the trend of two variables as the other variables change.
Covariance matrices are cases where the covariance concept is extended to multiple variables. For a data set with n variables, the covariance matrix is oneWherein each element represents the covariance between a pair of variables. The covariance matrix is generally expressed as:
；
In the covariance matrix, the diagonal elements are each variable and its own variance, and the non-diagonal elements are the covariances between the different variables. The covariance matrices of the measured data matrix a and the reference data matrix B are constructed in the above manner.
Further, the sample covariance matrix of the constructed time sequence sample data is subjected to feature decomposition, a feature value and a feature vector corresponding to the feature value are obtained after the feature decomposition, and key nodes of the dynamic network are screened according to element ranking in the feature vector corresponding to the maximum feature value.Representing the variableA sample covariance matrix consisting of N observations, which can be used as eigenvaluesAnd corresponding feature vectorAnd (3) representing. Feature vectorThe covariance matrix can be decomposed into eigenvalues and eigenvectors corresponding to the eigenvalues, each eigenvector corresponds to a variable in the measured data matrix A and the reference data matrix B, n elements in the eigenvector corresponding to the largest eigenvalue are compared, the h elements with the top numerical ranking are selected as key nodes of the dynamic network, and the measured data matrix A and the reference data matrix B respectively have the key nodes of the corresponding dynamic network. Near critical transition periodWhen in generalThe elements of (a) become non-zero and the top-ranked values show a more dramatic change, so that features with strong correlation to system state transitions can be reflected, and the elements in the feature vector corresponding to the largest feature value are selected to screen the key nodes of the dynamic network.
S4, calculating quantization indexes of the dynamic network of the actual measurement data matrix and the reference data matrix in each time window based on key nodes of the dynamic network respectively, calculating a system evaluation index in each time window according to the quantization indexes of the dynamic network of the actual measurement data matrix and the reference data matrix in each time window, generating early warning signals based on the system evaluation indexes and the key nodes of the corresponding dynamic network, and determining fault types.
In a specific embodiment, calculating, based on key nodes of the dynamic network, quantization indexes of the dynamic network of the actual measurement data matrix and the reference data matrix in each time window respectively, and calculating, according to the quantization indexes of the dynamic network of the actual measurement data matrix and the reference data matrix in each time window, a system evaluation index in each time window specifically includes:
Respectively calculating quantization indexes of the dynamic network of the actual measurement data matrix in a time window t by adopting the following steps of And a quantization index/>, with reference to the data matrix, of the dynamic network over a time window t：
；
；
Wherein,The element of the ith row and the jth column of the covariance matrix characteristic of the actual measurement data matrix at the time window t is h A, which is the first h A elements in the feature vector corresponding to the maximum feature value obtained after the covariance matrix characteristic of the actual measurement data matrix is decomposed, and the set of index marks of the key nodes of the corresponding dynamic network is，The element of the ith row and the jth column of the covariance matrix feature of the reference data matrix at the time window t is h B, which is the first h B elements in the feature vector corresponding to the maximum feature value obtained by decomposing the covariance matrix feature of the reference data matrix, and the set of index marks of the key nodes of the corresponding dynamic network is；
Calculating a system evaluation index of the time window t by adopting the following steps:
；
Wherein, Representing a system evaluation index; /(I)Covariance matrix use element set/>, which is measured data matrixCalculating quantization index of dynamic network in time window t, element setIs randomly selected from all elements in the eigenvector corresponding to the maximum eigenvalue of the covariance matrix of the measured data matrix, and the element setNumber and number of (2)The number of the sets is consistent,Covariance matrix usage element set/>, which is a reference data matrixCalculating quantization index of dynamic network in time window t, element setIs randomly selected from all elements in the eigenvector corresponding to the maximum eigenvalue of the covariance matrix of the reference data matrix, and the element setNumber andThe number of the sets is consistent.
In a specific embodiment, generating an early warning signal and determining a fault type based on the system evaluation index and the key node of the corresponding dynamic network specifically includes:
in response to determining that the system evaluation index is greater than the threshold, generating an early warning signal according to the corresponding time window Analyzing the key nodes corresponding to the index identifiers in the set to obtain corresponding fault types;
In response to determining that the system evaluation index is less than or equal to the threshold, the electromechanical equipment is determined to be in a healthy state.
Specifically, the quantization index of the dynamic network of the measured data matrix in the time window t is calculated on the basis of the key node of the dynamic network corresponding to the measured data matrix A and the key node of the dynamic network corresponding to the reference data matrix B respectivelyAnd a quantization index/>, with reference to the data matrix, of the dynamic network over a time window t. According toAndComputing System evaluation indexJudging whether a fault occurs based on the set standard threshold value, and identifying the fault type according to the key node. In particular embodiments, the threshold may be set as a threshold based on experimentation to maximize results of the electromechanical equipment calculation at the state of health. If the system evaluation index of a certain time window exceeds a threshold value, an early warning signal is sent out, and according to the moment/>, the system evaluation index of the certain time window exceeds the threshold valueAnd identifying the fault type and the cause by the key nodes corresponding to the index identifiers in the set. /(I)
The above steps S1-S4 do not necessarily represent the order between steps, but the step symbols indicate that the order between steps is adjustable.
The technical scheme of the embodiment of the application is further described by specific examples.
Example 1
Example 1 of the present application corresponds to the normal case. And analyzing the actual measurement historical data of the wind turbine generator in the health state. In order to verify the performance effect of the early-stage fault early-warning and identifying method for the electromechanical equipment provided with the SCADA system provided by the embodiment of the application under different running environments, three sections of measured historical data of the wind turbine generator in a healthy state are selected in the embodiment, and the early-stage fault early-warning and identifying method for the electromechanical equipment provided with the SCADA system provided by the embodiment of the application is used for test analysis. The test data are from a certain wind farm in northeast respectively:
1. No. 1A15 wind motor monitors data in 10 months of 2012 and 2 days of 0:00-3:00, and (180) samples in three hours, and the environmental wind speed in the period is kept at 3.4-5.4 m/s, so that the 3-level breeze is achieved.
2. Sample monitoring data of the No. 1A11 wind motor in 2012, 4 months and 4 days is 4:00-7:00, and the environmental wind speed in the period is kept at 10.8-13.8 m/s, so that 6-level strong wind is achieved.
3. Sample monitoring data of the No. 1A11 wind motor in 2012, 4, 7, 14:00-17:00, and three hours, wherein the environmental wind speed in the period is kept at 17.2-20.7 m/s, and the 8-level high wind is achieved.
The test results at different wind speeds are shown in figures 4-6. System evaluation index of system in breeze conditionThe system evaluation index/>, which is kept floating up and down at 0, of the system under the strong wind conditionCan be kept within 6; in case of heavy wind, the system evaluates the indexThe fluctuation is slightly larger but stabilized within 10. The result shows that when the wind turbine generator is in a healthy state, the system evaluation index/>, although the running environments of the three periods are differentThe fluctuations of (2) may be substantially stabilized within 10. Thus, the system evaluation indexIf the threshold of (2) is 10, if the system evaluates the indexIf the threshold value is exceeded, the wind turbine is in a critical working state, and an early warning signal can be sent out; if the system evaluates the indexAnd if the threshold value is not exceeded, the fan is in a stable running state.
Example 2
Embodiment 2 of the application corresponds to a gear box system failure of a wind turbine. The wind turbine generator system is shut down unplanned due to 41-component faults in 2012 of 1 month and 22 days. The maintenance personnel confirm that the fault point is that the function of the input template of the drive control system controller of the WT is invalid, the fault point is judged to be a fault of the gearbox, and the subsequent maintenance and element replacement result in shutdown for about 9 hours. In the embodiment, a section of actual measurement historical data of wind turbine generator set before and after fault state monitoring is selected for testing, namely, relevant monitoring project data of 60 minutes before and after fault shutdown is used as the actual measurement historical data. The test results are shown in fig. 7. From fig. 7, it can be seen that the system evaluation index of the dynamic networkAnd starting to steadily increase from the 1 st time window, exceeding a set threshold value when the 5 th time window, indicating that the system is in a critical state, and sending out an early warning signal. The peak is reached at time window 8 and 6 time windows remain after the threshold is exceeded. The fault alarm occurs in time window 13, after which a fault shutdown state is entered, and the indexReturning to the steady state, and keeping the state around 0. The verification result shows that the index is evaluated in the systemUnder the condition that the threshold value of (1) is set to 10, the early-stage fault early-warning and identifying method for the electromechanical equipment provided with the SCADA system can send out early-warning signals in 8 time windows (about 40 minutes) before faults occur. Therefore, the indexThe change of the fault can be used as an index for indicating the occurrence of the fault in advance, so that early warning is realized. And further analyzing the feature vector corresponding to the maximum feature value of the fault early-warning point. According to the dynamic network marker theory based on the sample covariance matrix, the elements in the feature vector corresponding to the maximum feature value can correspond to the original monitoring index, the elements in the feature vector are arranged in a descending order, and the elements with the top ranking can be selected as key nodes of the dynamic network. As shown in fig. 8, wherein the monitored variables: the gearbox input shaft 1 temperature, gearbox input shaft 2 temperature, gearbox oil temperature, and main shaft gearbox side bearing temperature become top-ranked 4 variables that can be selected as key nodes of the dynamic network. From the fault analysis, it can be obtained that the key nodes of the dynamic network formed in the 5 th time window can correctly indicate that the fan gearbox system has faults and are consistent with the overhauling result. Therefore, the early-fault early-warning and identifying method for the electromechanical equipment provided with the SCADA system provided by the embodiment of the application can comprehensively grasp the evolution process of the running state of the wind turbine generator, can keep stronger robustness in various complex running environments of the electromechanical equipment, timely detects the critical point of the health state degradation of the electromechanical equipment, and finally plays a role in timely sending an early-warning signal when the early defect of the electromechanical equipment occurs and avoiding accidents.
With further reference to fig. 9, as an implementation of the method shown in the foregoing drawings, the present application provides an embodiment of an early-stage fault early-warning and identifying device for electromechanical equipment equipped with a SCADA system, where the embodiment of the device corresponds to the embodiment of the method shown in fig. 2, and the device may be applied to various electronic devices specifically.
The embodiment of the application provides an early-stage fault early-warning and identifying device of electromechanical equipment provided with a SCADA system, which comprises the following components:
The dynamic network construction module 1 is configured to select variables closely related to the operation state and having complex coupling relation from all main monitoring items of the SCADA system of the electromechanical equipment, and map the variables into nodes of the dynamic network;
The theoretical prediction module 2 is configured to construct a health state theoretical prediction model for each node of the dynamic network, acquire actual measurement historical data of each node acquired by the SCADA system in each sampling time in the previous sampling period when the electromechanical equipment is in a health state, construct an actual measurement data matrix of the previous sampling period, and input the actual measurement data matrix of the previous sampling period into the health state theoretical prediction model to obtain a reference data matrix of the current sampling period;
The key node screening module 3 is configured to respectively construct a covariance matrix of the actual measurement data matrix and a covariance matrix of the reference data matrix of each time window divided in the current sampling period, respectively perform characteristic decomposition, calculate a characteristic value obtained after the decomposition and a characteristic vector corresponding to the characteristic value, and screen key nodes of the dynamic network according to the characteristic vector corresponding to the maximum characteristic value and element ranking of the characteristic vector;
The fault analysis module 4 is configured to calculate quantization indexes of the dynamic network of the actual measurement data matrix and the reference data matrix in each time window based on the key nodes of the dynamic network respectively, calculate a system evaluation index in each time window according to the quantization indexes of the dynamic network of the actual measurement data matrix and the reference data matrix in each time window, generate early warning signals based on the system evaluation indexes and the key nodes of the corresponding dynamic network, and determine fault types.
Referring now to fig. 10, there is illustrated a schematic diagram of a computer apparatus 1000 suitable for use in an electronic device (e.g., a server or terminal device as illustrated in fig. 1) for implementing an embodiment of the present application. The electronic device shown in fig. 10 is merely an example, and should not impose any limitation on the functionality and scope of use of embodiments of the present application.
As shown in fig. 10, the computer apparatus 1000 includes a Central Processing Unit (CPU) 1001 and a Graphics Processor (GPU) 1002, which can perform various appropriate actions and processes according to a program stored in a Read Only Memory (ROM) 1003 or a program loaded from a storage section 1009 into a Random Access Memory (RAM) 1004. In the RAM 1004, various programs and data required for the operation of the computer device 1000 are also stored. The CPU 1001, the GPU1002, the ROM 1003, and the RAM 1004 are connected to each other by a bus 1005. An input/output (I/O) interface 1006 is also connected to bus 1005.
The following components are connected to the I/O interface 1006: an input section 1007 including a keyboard, a mouse, and the like; an output portion 1008 including a speaker, such as a Liquid Crystal Display (LCD), or the like; a storage section 1009 including a hard disk or the like; and a communication section 1010 including a network interface card such as a LAN card, a modem, or the like. The communication section 1010 performs communication processing via a network such as the internet. The drive 1011 may also be connected to the I/O interface 1006 as needed. A removable medium 1012 such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, or the like is installed on the drive 1011 as necessary, so that a computer program read out therefrom is installed into the storage section 1009 as necessary.
In particular, according to embodiments of the present disclosure, the processes described above with reference to flowcharts may be implemented as computer software programs. For example, embodiments of the present disclosure include a computer program product comprising a computer program embodied on a computer readable medium, the computer program comprising program code for performing the method shown in the flowcharts. In such embodiments, the computer program may be downloaded and installed from a network via the communications portion 1010, and/or installed from the removable media 1012. The above-described functions defined in the method of the present application are performed when the computer program is executed by a Central Processing Unit (CPU) 1001 and a Graphics Processor (GPU) 1002.
It should be noted that the computer readable medium according to the present application may be a computer readable signal medium or a computer readable medium, or any combination of the two. The computer readable medium can be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor apparatus, device, or means, or a combination of any of the foregoing. More specific examples of the computer-readable medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In the context of this document, a computer readable medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution apparatus, device, or apparatus. In the present application, however, the computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, with the computer-readable program code embodied therein. Such a propagated data signal may take any of a variety of forms, including, but not limited to, electro-magnetic, optical, or any suitable combination of the foregoing. A computer readable signal medium may be any computer readable medium that is not a computer readable medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution apparatus, device, or apparatus. Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to: wireless, wire, fiber optic cable, RF, etc., or any suitable combination of the foregoing.
Computer program code for carrying out operations of the present application may be written in any combination of one or more programming languages, including an object oriented programming language such as Java, smalltalk, C ++ and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any kind of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or may be connected to an external computer (for example, through the Internet using an Internet service provider).
The flowcharts and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of apparatus, methods and computer program products according to various embodiments of the present application. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based devices which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.
The modules involved in the embodiments of the present application may be implemented in software or in hardware. The described modules may also be provided in a processor.
As another aspect, the present application also provides a computer-readable medium that may be contained in the electronic device described in the above embodiment; or may exist alone without being incorporated into the electronic device. The computer readable medium carries one or more programs which, when executed by the electronic device, cause the electronic device to: selecting variables closely related to the operation state and having complex coupling relation from all main monitoring items of an SCADA system of the electromechanical equipment, and mapping the variables into nodes of a dynamic network; constructing a health state theoretical prediction model aiming at each node of a dynamic network, acquiring actual measurement historical data of each node acquired by an SCADA system in each sampling time in a previous sampling period when electromechanical equipment is in a health state, constructing an actual measurement data matrix in the previous sampling period, and inputting the actual measurement data matrix in the previous sampling period into the health state theoretical prediction model to obtain a reference data matrix in the current sampling period; respectively constructing covariance matrixes of measured data matrixes and covariance matrixes of reference data matrixes of each time window divided in a current sampling period, respectively performing feature decomposition, calculating feature values and corresponding feature vectors obtained after the decomposition, and screening key nodes of a dynamic network according to the feature vectors corresponding to the maximum feature values and element ranks thereof; the method comprises the steps of respectively calculating quantization indexes of a dynamic network of an actual measurement data matrix and a reference data matrix in each time window based on key nodes of the dynamic network, calculating a system evaluation index in each time window according to the quantization indexes of the dynamic network of the actual measurement data matrix and the reference data matrix in each time window, generating early warning signals based on the system evaluation indexes and the key nodes of the corresponding dynamic network, and determining fault types.
The above description is only illustrative of the preferred embodiments of the present application and of the principles of the technology employed. It will be appreciated by persons skilled in the art that the scope of the application referred to in the present application is not limited to the specific combinations of the technical features described above, but also covers other technical features formed by any combination of the technical features described above or their equivalents without departing from the inventive concept described above. Such as the above-mentioned features and the technical features disclosed in the present application (but not limited to) having similar functions are replaced with each other.
Claims (7)
1. An early-stage fault early-warning and identifying method for electromechanical equipment provided with a SCADA system is characterized by comprising the following steps:
Selecting variables closely related to the operation state and having complex coupling relation from all main monitoring items of an SCADA system of the electromechanical equipment, and mapping the variables into nodes of a dynamic network;
Constructing a health state theoretical prediction model aiming at each node of the dynamic network, acquiring actual measurement historical data of each node acquired by an SCADA system in a previous sampling period when the electromechanical equipment is in a health state, constructing an actual measurement data matrix of the previous sampling period, and inputting the actual measurement data matrix of the previous sampling period into the health state theoretical prediction model to obtain a reference data matrix of the current sampling period;
Respectively constructing covariance matrixes of the actual measurement data matrixes and covariance matrixes of the reference data matrixes of each time window divided in the current sampling period, respectively carrying out feature decomposition, calculating feature values obtained after the decomposition and corresponding feature vectors, and screening key nodes of the dynamic network according to the feature vectors corresponding to the maximum feature values and element ranks thereof, wherein the key nodes comprise the following specific steps:
The covariance matrix of the measured data matrix and the covariance matrix of the reference data matrix can both represent variables Sample covariance matrix/>, consisting of N observationsAnd carrying out characteristic decomposition on the sample covariance matrix, wherein the characteristic decomposition is shown in the following formula:
；
Wherein, Is a diagonal matrix containing eigenvalues/>, of the sample covariance matrix C；Is an orthogonal matrix comprising eigenvectors/>, of a sample covariance matrix C，；
The elements of the ith row and jth column of the sample covariance matrix at time window t are represented as:
；
Wherein, ，Is the variable/>, in the sample covariance matrix CAndCovariance,Representing the maximum eigenvalue,AndVariable/>, respectivelyAndFeature vector/>, corresponding to maximum feature value, in sample covariance matrix of time window tAn ith element and a jth element; comparing the eigenvector/>, corresponding to the maximum eigenvalueSelecting variables corresponding to the h elements with the top numerical ranking as key nodes of the dynamic network, wherein the set of index marks is defined asThe value of h is set to satisfyAndWhereinRepresenting the variableThe ith element of the eigenvector corresponding to the largest eigenvalue in the sample covariance matrix of the critical transition time window t', the P value is a positive real number of [0,1 ]IsAn upper bound on the number of all elements in a system;
Calculating the quantization index of the dynamic network of the actual measurement data matrix and the reference data matrix in each time window based on the key nodes of the dynamic network respectively, and calculating the system evaluation index in each time window according to the quantization index of the dynamic network of the actual measurement data matrix and the reference data matrix in each time window, wherein the system evaluation index specifically comprises the following steps:
respectively calculating quantization indexes of the dynamic network of the measured data matrix in a time window t by adopting the following steps of And the quantization index/>, of the dynamic network of the reference data matrix over a time window t：
；
；
Wherein,The element of the ith row and the jth column of the covariance matrix feature of the actual measurement data matrix at the time window t is h A, which is the first h A elements in the feature vector corresponding to the maximum feature value obtained after the covariance matrix feature of the actual measurement data matrix is decomposed, and the set of index marks of the key nodes of the corresponding dynamic network is，The element of the ith row and the jth column of the covariance matrix characteristic of the reference data matrix at the time window t is h B, which is the first h B elements in the eigenvector corresponding to the maximum eigenvalue obtained after the covariance matrix characteristic of the reference data matrix is decomposed, and the set of index marks of the key nodes of the corresponding dynamic network is；
Calculating a system evaluation index of the time window t by adopting the following steps:
；
Wherein, Representing a system evaluation index; /(I)Covariance matrix use element which is the measured data matrixSet calculation of quantization index of dynamic network in time window t, element setIs randomly selected from all elements in the eigenvector corresponding to the maximum eigenvalue of the covariance matrix of the measured data matrix, and the element setNumber and number of (2)The number of the sets is consistent,Covariance matrix use element set/>, which is the reference data matrixCalculating quantization index of dynamic network in time window t, element setIs randomly selected from all elements in the eigenvector corresponding to the maximum eigenvalue of the covariance matrix of the reference data matrix, and the element setNumber andThe number of the sets is consistent; generating an early warning signal and determining a fault type based on the system evaluation index and the key node of the dynamic network corresponding to the system evaluation index, wherein the method specifically comprises the following steps:
In response to determining that the system evaluation index is greater than a threshold, generating the early warning signal according to a corresponding time window Analyzing the key nodes corresponding to the index identifiers in the set to obtain corresponding fault types;
In response to determining that the system evaluation index is less than or equal to a threshold, the electromechanical equipment is determined to be in a healthy state.
2. The method for early warning and identifying of an early failure of an electromechanical installation equipped with a SCADA system of claim 1, wherein the theoretical state of health prediction model comprises a trained long and short term memory neural network.
3. The early-fault pre-warning and identifying method for electromechanical equipment provided with a SCADA system according to claim 1, wherein the actual measurement history data is normalized data.
4. The early-fault warning and identifying method for electromechanical equipment equipped with SCADA system according to claim 1, wherein the covariance matrix of the measured data matrix is expressed as:
；
The covariance matrix of the reference data matrix is expressed as:
；
Wherein, AndThe elements in the measured data matrix A and the reference data matrix B are respectively represented, and the corresponding values are covariance of the ith and jth variables.
5. An early-fault early-warning and identifying device for electromechanical equipment provided with a SCADA system, which is characterized by comprising the following components:
the dynamic network construction module is configured to select variables which are closely related to the operation state and have complex coupling relations from all main monitoring items of the SCADA system of the electromechanical equipment, and map the variables into nodes of the dynamic network;
The theory prediction module is configured to construct a health state theory prediction model for each node of the dynamic network, acquire actual measurement historical data of each node acquired by the SCADA system in each sampling time in the previous sampling period when the electromechanical equipment is in a health state, construct an actual measurement data matrix of the previous sampling period, and input the actual measurement data matrix of the previous sampling period into the health state theory prediction model to obtain a reference data matrix of the current sampling period;
The key node screening module is configured to respectively construct covariance matrixes of the actual measurement data matrixes and covariance matrixes of the reference data matrixes of each time window divided in the current sampling period, respectively perform feature decomposition, calculate feature values obtained after the decomposition and corresponding feature vectors, and screen key nodes of the dynamic network according to the feature vectors corresponding to the maximum feature values and element ranks thereof, and specifically comprises the following steps:
The covariance matrix of the measured data matrix and the covariance matrix of the reference data matrix can both represent variables Sample covariance matrix/>, consisting of N observationsAnd carrying out characteristic decomposition on the sample covariance matrix, wherein the characteristic decomposition is shown in the following formula:
；
Wherein, Is a diagonal matrix containing eigenvalues/>, of the sample covariance matrix C；Is an orthogonal matrix comprising eigenvectors/>, of a sample covariance matrix C，；
The elements of the ith row and jth column of the sample covariance matrix at time window t are represented as:
；
Wherein, ，Is the variable/>, in the sample covariance matrix CAndCovariance,Representing the maximum eigenvalue,AndVariable/>, respectivelyAndFeature vector/>, corresponding to maximum feature value, in sample covariance matrix of time window tAn ith element and a jth element; comparing the eigenvector/>, corresponding to the maximum eigenvalueSelecting variables corresponding to the h elements with the top numerical ranking as key nodes of the dynamic network, wherein the set of index marks is defined asThe value of h is set to satisfyAndWhereinRepresenting the variableThe ith element of the eigenvector corresponding to the largest eigenvalue in the sample covariance matrix of the critical transition time window t', the P value is a positive real number of [0,1 ]IsAn upper bound on the number of all elements in a system;
The fault analysis module is configured to calculate quantization indexes of the dynamic network of the actual measurement data matrix and the reference data matrix in each time window based on key nodes of the dynamic network respectively, and calculate a system evaluation index in each time window according to the quantization indexes of the dynamic network of the actual measurement data matrix and the reference data matrix in each time window, and specifically comprises the following steps:
respectively calculating quantization indexes of the dynamic network of the measured data matrix in a time window t by adopting the following steps of And the quantization index/>, of the dynamic network of the reference data matrix over a time window t：
；
；
Wherein,The element of the ith row and the jth column of the covariance matrix feature of the actual measurement data matrix at the time window t is h A, which is the first h A elements in the feature vector corresponding to the maximum feature value obtained after the covariance matrix feature of the actual measurement data matrix is decomposed, and the set of index marks of the key nodes of the corresponding dynamic network is，The element of the ith row and the jth column of the covariance matrix characteristic of the reference data matrix at the time window t is h B, which is the first h B elements in the eigenvector corresponding to the maximum eigenvalue obtained after the covariance matrix characteristic of the reference data matrix is decomposed, and the set of index marks of the key nodes of the corresponding dynamic network is；
Calculating a system evaluation index of the time window t by adopting the following steps:
；
Wherein, Representing a system evaluation index; /(I)Covariance matrix use element which is the measured data matrixSet calculation of quantization index of dynamic network in time window t, element setIs randomly selected from all elements in the eigenvector corresponding to the maximum eigenvalue of the covariance matrix of the measured data matrix, and the element setNumber and number of (2)The number of the sets is consistent,Covariance matrix use element set/>, which is the reference data matrixCalculating quantization index of dynamic network in time window t, element setIs randomly selected from all elements in the eigenvector corresponding to the maximum eigenvalue of the covariance matrix of the reference data matrix, and the element setNumber andThe number of the sets is consistent; generating an early warning signal and determining a fault type based on the system evaluation index and the key node of the dynamic network corresponding to the system evaluation index, wherein the method specifically comprises the following steps:
In response to determining that the system evaluation index is greater than a threshold, generating the early warning signal according to a corresponding time window Analyzing the key nodes corresponding to the index identifiers in the set to obtain corresponding fault types;
In response to determining that the system evaluation index is less than or equal to a threshold, the electromechanical equipment is determined to be in a healthy state.
6. An electronic device, comprising:
One or more processors;
Storage means for storing one or more programs,
When executed by the one or more processors, causes the one or more processors to implement the method of any of claims 1-4.
7. A computer readable storage medium, on which a computer program is stored, characterized in that the program, when being executed by a processor, implements the method according to any of claims 1-4.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410400068.2A CN117992859B (en) | 2024-04-03 | 2024-04-03 | Early-stage fault early-warning and identifying method and device for electromechanical equipment provided with SCADA system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410400068.2A CN117992859B (en) | 2024-04-03 | 2024-04-03 | Early-stage fault early-warning and identifying method and device for electromechanical equipment provided with SCADA system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117992859A CN117992859A (en) | 2024-05-07 |
| CN117992859B true CN117992859B (en) | 2024-06-07 |
Family
ID=90900942
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410400068.2A Active CN117992859B (en) | 2024-04-03 | 2024-04-03 | Early-stage fault early-warning and identifying method and device for electromechanical equipment provided with SCADA system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117992859B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN118134466B (en) * | 2024-05-08 | 2024-07-09 | 国网上海市电力公司 | Substation equipment fault troubleshooting method based on service scene |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2013061853A (en) * | 2011-09-14 | 2013-04-04 | Toshiba Corp | Process monitoring/diagnosis support device |
| CN114139638A (en) * | 2021-12-03 | 2022-03-04 | 中国电建集团贵州电力设计研究院有限公司 | Fan blade icing fault diagnosis method considering multivariable correlation |
| CN116383636A (en) * | 2023-03-30 | 2023-07-04 | 武汉理工大学 | Coal mill fault early warning method based on PCA and LSTM fusion algorithm |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP6763910B2 (en) * | 2018-05-30 | 2020-09-30 | 横河電機株式会社 | Anomaly detection device, anomaly detection method, anomaly detection program and recording medium |
-
2024
- 2024-04-03 CN CN202410400068.2A patent/CN117992859B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2013061853A (en) * | 2011-09-14 | 2013-04-04 | Toshiba Corp | Process monitoring/diagnosis support device |
| CN114139638A (en) * | 2021-12-03 | 2022-03-04 | 中国电建集团贵州电力设计研究院有限公司 | Fan blade icing fault diagnosis method considering multivariable correlation |
| CN116383636A (en) * | 2023-03-30 | 2023-07-04 | 武汉理工大学 | Coal mill fault early warning method based on PCA and LSTM fusion algorithm |
Non-Patent Citations (1)
| Title |
|---|
| 基于神经网络的风力机偏航系统故障诊断技术研究;邓子豪;《中国优秀硕士学位论文全文数据库 工程科技Ⅱ辑》;20230115;全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117992859A (en) | 2024-05-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Shamayleh et al. | IoT based predictive maintenance management of medical equipment | |
| CN117992859B (en) | Early-stage fault early-warning and identifying method and device for electromechanical equipment provided with SCADA system | |
| US11835030B2 (en) | Methods and systems for predicting risk of observable damage in wind turbine gearbox components | |
| Tsui et al. | Prognostics and health management: A review on data driven approaches | |
| Li et al. | Multi-sensor data-driven remaining useful life prediction of semi-observable systems | |
| Haghshenas et al. | Predictive digital twin for offshore wind farms | |
| US20130060524A1 (en) | Machine Anomaly Detection and Diagnosis Incorporating Operational Data | |
| Welte et al. | Models for lifetime estimation: an overview with focus on applications to wind turbines | |
| JP2006500694A (en) | Apparatus and method for monitoring technical equipment including a plurality of systems, in particular power plant equipment | |
| JP2022534070A (en) | Fault prediction using gradient-based sensor identification | |
| Mathew et al. | Regression kernel for prognostics with support vector machines | |
| CN114254904B (en) | Method and device for evaluating operation health degree of engine room of wind turbine generator | |
| Liu et al. | A hybrid multi-stage methodology for remaining useful life prediction of control system: Subsea Christmas tree as a case study | |
| CN117422447A (en) | Transformer maintenance strategy generation method, system, electronic equipment and storage medium | |
| CN117171657A (en) | Wind power generation equipment fault diagnosis method and device, electronic equipment and storage medium | |
| CN118035731A (en) | Electricity safety monitoring and early warning method and service system | |
| Hu et al. | Mutual information-based feature disentangled network for anomaly detection under variable working conditions | |
| Akcan et al. | A new approach for remaining useful life prediction of bearings using 1D-ternary patterns with LSTM | |
| CN116467575A (en) | Real-time data monitoring system and method design of ball screw pair system | |
| CN115456041A (en) | Equipment fault early warning method and device, computing equipment and storage medium | |
| Peng et al. | Regime-Switching Model With Adaptive Adjustments for Degradation Prognosis | |
| CN115456168A (en) | Training method and energy consumption determination method and device for reinforcement learning model | |
| Abdulrahim et al. | IoT Based Predictive Maintenance Management of Medical Equipment | |
| Agrell et al. | Pitfalls of machine learning for tail events in high risk environments | |
| CN117992762B (en) | Overheat early warning method and device for stator winding of water-cooled steam turbine generator |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant |